Performance of a water reclamation plant using ultrafiltration (UF) and reverse osmosis (RO) treating 280,000 m3/d of wastewater was evaluated over 1 year. Statistical analyses were performed on flow rate, temperature, pH, total suspended solids (TSS), total dissolved solids (TDS), biochemical oxygen demand (BOD5) and total coliforms. Variations in flow rates coincided with those in temperature, both being seasonal, but plant performance was not highly influenced by such variations. The RO system recovered 85% of water flow. Data on process variables conform to a normal probability distribution and reveal the high process efficiency and reliability of UF and RO systems. Plant efficiencies were >99% for TSS, TDS, BOD5 and total coliforms. Efficiencies were the highest for TSS and total coliforms in the UF system, while they were the highest for TDS and BOD5 in the RO system. Cumulative frequency distribution analyses indicate that RO plays an important role in maintaining a stable plant performance and high quality reclaimed water. The UF system proved essential for complimenting successful performance of the RO system. Reclaimed water satisfies, at 99% frequency, the quality standards for potable water concerning TSS, TDS, BOD5 and total coliforms even though membranes have been operating continuously for 6 consecutive years.

INTRODUCTION

Water reclamation is expected to increase in future years, mainly due to a steady increase of the world's population and water use contrasted with a decrease in the availability of freshwater supply and in the affordability of meeting such an increase (Angelakis et al. 2003; Asano et al. 2007). This will require a coordinated effort involving water management and purification, water conservation and water reuse (Goosen & Shayya 1999). Reclamation of treated wastewater from municipalities and industries has been considered a supplementary source of water supply in many parts of the world, particularly in some areas where water resources are scarce and population and economic growth is rapid (Hamoda 2004).

Membrane processes have been considered a promising technology for wastewater recycling and reuse. Numerous membrane technologies based on micro-filtration, ultrafiltration (UF), nanofiltration and reverse osmosis (RO) have been used downstream of biological treatment units to act as selective barriers that permit separation and removal of contaminants in a fluid by a combination of sieving and sorption diffusion mechanisms (Asano et al. 2007; Chang et al. 2007; Lazarova et al. 2013).

RO is a membrane-based process that employs semi-permeable membranes under high pressure differentials to separate and remove dissolved solids, organics, submicron colloidal matter, color, nitrate, phosphorus and bacteria from wastewater (Asano et al. 2007). The RO process has been used extensively in reclamation of wastewater and in desalination of brackish/sea water. Removal of organic contaminants by RO was first demonstrated by Hamoda et al. (1973), and was followed by numerous studies investigating the use of RO for the removal of individual organics from domestic and industrial wastewaters (Peng et al. 2004; Bodalo-Santoyo et al. 2004; Arsuaga et al. 2006; Nataraj et al. 2006; Dolar et al. 2013).

UF membrane processes use a wide range of pore size and an operating pressure range of 3–10 bar (Cheryan 1986). The separation mechanism in UF is selective sieving through the membrane pores. Dissolved salts and small organic molecules pass through the semi-permeable membrane in the liquid phase while larger molecules and solids are rejected and concentrated in the retentate (Tansel et al. 1995). UF is often used as a pre-treatment process for RO to reduce the adverse effects on downstream RO. In a previous study, Glueckstern et al. (2000) used two-stage media filtration and UF as pre-treatment for RO to treat contaminated brackish fish-pond effluents and showed that UF was more effective than multi-media filtration in improving water quality. In another study, a significant reduction in total solids concentration and chemical oxygen demand (COD) was achieved by the UF system treating poultry processing wastewater (Lo et al. 2005). Furthermore, Nelson (2006) reported that UF reduced organic and microbial pollutants from poultry processing wastewater while Avula et al. (2009) found that UF improved the quality of recycled poultry wastewater and provided a solution to water resources limitations.

Combinations of UF and RO technologies have also been investigated. In one study (Fababuj-Roger et al. 2007), tannery wastewater was reclaimed by UF and RO after a conventional physical–chemical treatment. The results showed that a simple combination of physical–chemical treatment and UF was not sufficient to remove the soluble COD from wastewater, leaving a concentration more than 2,000 mg/L in the permeates from all tested UF membranes. Conversely, more than 98% rejection of both salts and COD were achieved when RO treatment was included. Their study clearly demonstrates the capability of RO systems, especially when the effluent contains high organic matter content. In another study (Tomaszewska et al. 2005), oily wastewater and bilge water were treated using a combination of UF and RO systems operating via two-stage treatment, with UF being used in the first stage and RO in the second stage. It was found that permeate from the first stage had oil content below 10 ppm, was free of suspended solids and almost all turbidities were removed. In the second stage, total organic carbon (TOC) removal was more than 70% while 90% of all cations examined (Na+, K+ , Mg2+, Ca2+, Zn2+, Mn2+, Al3+, Li+), P2O5 and sulfate were removed. In addition, Murthy & Choudhari (2009) investigated the application of UF and RO membranes in a pilot plant treating wastewater for removal of color and contaminants. The results showed that the rejection efficiencies of UF and RO membranes for total dissolved solids (TDS), COD, biochemical oxygen demand (BOD5), sulfate and potassium were 97.9%, 96.8%, 97.9%, 99.7% and 94.65%, respectively. Similar results were reported in other studies (Shi & Benjamin 2009; Huang et al. 2011).

Previous studies on UF and RO, as reported in the literature, have been conducted mostly on laboratory or pilot-scale systems but not on full-scale operations. Therefore, this study was carried out to evaluate the performances of UF and RO systems at full-scale operation, to give a clearer understanding of the performance at a large scale. The Sulaibiya wastewater reclamation plant in the State of Kuwait, being the largest UF/RO plant in the world, was investigated in this study to determine the process efficiency and performance reliability of the UF and RO membrane systems.

MATERIALS AND METHODS

Plant description

The Sulaibiya wastewater treatment plant (WWTP) is located at Sulaibiya near Kuwait City. It was commissioned in the year 2005, with a design capacity of 375,000 m3/day which was increased to 425,000 m3/day, as the world's largest membrane-based water reclamation facility utilizing UF and RO systems. The reclaimed water from this facility is used for several non-potable uses. The water reclamation facility (Figure 1) receives secondary-treated municipal wastewater, which is pre-filtered with disk filters and then fed to the UF system. The UF permeate feeds the RO plant, and UF retentate is recycled to the WWTP. The system receives almost 100% of the effluent from the biological treatment plant since UF retentate is recycled. RO is placed just after UF and therefore, the inlet feed of RO is also 375,000 m3/day. The RO system is constructed for 85% water recovery and thus, the production rate is expected to be 318,750 m3/day. The WWTP is designed to produce an effluent with an average monthly value of less than 20 mg/L BOD5 and 20 mg/L total suspended solids (TSS). It should be noted that the effluent TDS from this plant has to be <100 mg/L (the influent contains 1,280 mg/L), which is apparently better than World Health Organization (WHO 2008) potable water guidelines of <1,000 mg/L. As shown in Figure 1, the preliminary treatment at Ardiya consists of particulate and grit removal, as well as fat, oil and grease removal. The wastewater is then pumped for about 25 km to Sulaibiya for secondary and advanced (reclamation) treatment stages. To enhance biological removal of nitrogen and phosphorus in the WWTP, anoxic and aerobic systems are used in the secondary treatment stage in addition to secondary clarifiers. To minimize variation in flow, buffer volume was taken into account in the design of the facilities at Ardiya, and the aeration basins and the clarifiers at Sulaibiya. For sludge treatment, both aerobic digesters and drying beds are used, which are suitable under warm to hot climatic conditions prevailing in Kuwait. According to the regulatory agencies in the State of Kuwait, the digested sludge could be used as organic fertilizer or soil conditioner. The reclaimed water could be used for non-potable uses such as irrigation and for groundwater recharge.

Figure 1

Flow diagram of treatment processes at Sulaibiya WWTP.

Figure 1

Flow diagram of treatment processes at Sulaibiya WWTP.

UF system

Since RO systems require pre-treatment as to protect RO membranes from fouling, UF was selected to provide appropriate pre-treatment of the secondary-treated municipal effluent before being fed to the RO. The UF technology is robust, has favorable life cycle costs and provides better quality water to the RO membranes. The characteristics of the UF system used in this plant are presented in Table 1. Each UF unit can be operated individually. These units are regularly backwashed to ensure removal of suspended matter retained on the membranes. The backwash water is pumped back upstream of the WWTP to receive appropriate treatment and accomplish the maximum total water recovery for the plant. The influent to the UF first passes through a disk filter and subsequently, a small amount of coagulant (ferric chloride at 1–2 mg/L) is added to coagulate fine particulates and possibly allow some TOC removal to facilitate the operation of the plant. The silt density index of the UF product is consistently below 2, which is the key standard for the RO plant performance (Gagne 2002).

Table 1

Characteristics of membrane systems employed at Sulaibiya water reclamation plant

Membrane system Membrane type (manufacturer; catalogue no.) Membrane configuration Membrane arrangement Membrane area 
UF Polyvinyldene fluoride; (Norit X-Flow, The Netherlands; XIGA SXL-225) Capillary hydrophilic hollow fibers Membranes are packed in 20 × 152 cm membrane elements (35 m2/element), 4 membrane elements are placed inside a membrane housing. There are 68 skids, each with 32 membrane housings for a total of 8,704 membrane elements (4 × 32 × 68) 8,704 × 35 m2 = 304,640 m2 
RO Polyamide composite (Toray, USA; TML 20–400) Spiral wound Membrane modules of 42 identical skids in a 4:2:1 array of modules. Each module contains about 504 RO elements (72 pressure vessels × 7 RO element/vessel) for a total of 21,168 membrane elements (7 × 72 × 42) 21,168 × 37 m2 = 783,216 m2 
Membrane system Membrane type (manufacturer; catalogue no.) Membrane configuration Membrane arrangement Membrane area 
UF Polyvinyldene fluoride; (Norit X-Flow, The Netherlands; XIGA SXL-225) Capillary hydrophilic hollow fibers Membranes are packed in 20 × 152 cm membrane elements (35 m2/element), 4 membrane elements are placed inside a membrane housing. There are 68 skids, each with 32 membrane housings for a total of 8,704 membrane elements (4 × 32 × 68) 8,704 × 35 m2 = 304,640 m2 
RO Polyamide composite (Toray, USA; TML 20–400) Spiral wound Membrane modules of 42 identical skids in a 4:2:1 array of modules. Each module contains about 504 RO elements (72 pressure vessels × 7 RO element/vessel) for a total of 21,168 membrane elements (7 × 72 × 42) 21,168 × 37 m2 = 783,216 m2 

RO system

The characteristics of the RO system adopted in the Sulaibiya plant are shown in Table 1. This system is used to desalinate the wastewater effluent to 100 mg/L TDS and to provide an additional barrier to bacteria and viruses. The salinity of the secondary-treated effluent has an average monthly value of 1,280 mg/L TDS, with a maximum value of 3,014 mg/L. The RO system modules are arranged in a 4:2:1 array, forming three stages of RO treatment. The first stage recovers 50%, the second stage recovers 50% and the third stage recovers 40% of the flow. Recovery of water by the RO system is limited to 85% by calcium phosphate precipitation, which can be often a limiting factor for recovery of water in the membrane desalination systems in municipal wastewater (Gagne 2002). RO effluent passes through the stripper unit to remove the CO2 to adjust the pH with a minimum of caustic soda prior to distribution, then the product is chlorinated before leaving the plant. The system's brine is disposed of into the waters of the Arabian Gulf. The WANDA Control module is used extensively to verify and optimize the normal operating procedures, such as backwash operations and pump switches. To minimize fouling, the membranes are regularly cleaned in place using surfactants, sulfuric acid, biocide and sodium hydroxide.

Plant operation

The treatment steps have been selected to minimize the use of power and chemicals, and are simple to operate and maintain. A high degree of standby equipment and redundancy was incorporated into the design to ensure reliable operation. A key to the successful operation of this facility is to employ highly trained and motivated staff. Operating personnel include a number of managers, chemists and engineers, as well as qualified technicians and laborers. Training to the appropriate level was provided at the manufacturers' premises and on-site during commissioning and testing.

The plant receives primarily domestic wastewater with only slight variations in influent concentration levels. Moreover, the plant has several storage tanks to balance water flows of different treatment steps. This contributes to steady operation of the plant and low time variability. The characteristics of the raw wastewater are typical of those of high-strength sewage. A unique feature is the high TDS of the raw wastewater mainly due to the infiltration of saline water into the sewerage collection system as a result of seawater intrusion and subsurface water rise.

METHODOLOGY

This study used 12 months of complete daily measurements of quantitative data from Sulaibiya WWTP for the period 1 September 2010 to 31 August 2011 to build a database. Although the study is based on real data collected daily, the database suffers from some missing values over some weekends which can be compensated for by taking averages. In the studied WWTP, it is common to make daily analytical measurements of process variables on samples collected from influent and effluent of each treatment stage in general and each treatment process in the reclamation plant such as UF and RO in particular. The process variables considered for the database are flow rate and temperature whereas quantitative water quality data includes solids (measured as TSS and TDS), organic matter (measured as BOD5, and coliform bacteria (measured as total coliforms). Flow rate and temperature are recorded from on-line signals coming from sensors whereas the water quality parameters were determined according to Standard Methods (APHA 2005).

The final data set covers a homogeneous representative period of 320 consecutive days, excluding some weekends, where each day is considered as a new sample. Monthly averages for a period of 12 months were determined based on the daily sets of data. The data were subjected to statistical analysis, including basic statistical descriptors of the parameters included in the database, the normal probability distribution plots and the cumulative frequency distributions of the influent and effluent concentrations of the UF and RO systems. Such statistical methods for data analysis have become common tools for plant data (Asano et al. 2007; Mujeriego & Peters 2008). In this study, the main objective of data analysis was to statistically quantify the process efficiency and the performance reliability of both UF and RO membrane systems of the water reclamation plant.

RESULTS AND DISCUSSION

In the following section, plant performance will be analysed through: (1) percentage reduction efficiency and a comparison of the effluent concentrations with the water quality standards for reuse; and (2) cumulative frequency distributions of the influent and effluent concentrations. The statistical analysis was performed on the variables pH, TDS, TSS, BOD and total coliforms. Each of these five variables was monitored in both the influent and the effluent to the UF and the RO systems. Both the daily records and the monthly averages were used in the analysis.

Operating parameters

The performance of a treatment system is usually influenced by two important ‘input’ variables, namely the influent flow rate and temperature. The effluent variables considered in this study are pH, TSS, TDS, BOD and total coliforms. These five parameters were presented in 15 groups of data values, namely five groups for UF influent, five groups for UF effluent/RO influent and five groups for RO effluent. The first step in the statistical analyses conducted on plant performance data attempted to examine the statistical distribution of each variable. To apply the statistical tests associated with normal distribution, one of the variables (TDS) was examined. The normal probability plots for TDS in influent and effluent of RO are shown in Figure 2, confirming that data follow a normal distribution pattern. The regression lines obtained showed high correlation coefficients of 0.9182 and 0.9659 for influent and effluent TDS, respectively. Similarly, other variables inspected were normally distributed. Thus all the 15 groups of data conform to a normal probability distribution (regression lines). Such features make characterization of plant performance much simpler.

Figure 2

Normal probability plots of TDS for the RO system.

Figure 2

Normal probability plots of TDS for the RO system.

Figure 3 illustrates the daily variations in each of the flow rate and temperature over a period of 1 year. The water reclamation plant is currently operated under its design flow, with average influent flow rate of 280,000 ± 14,800 m3/d, as compared to the design flow of 425,000 m3/d. Seasonal variations in flow rate were minor (±5%), with highest flow rates being recorded in the summer months when water consumption is high, hence wastewater generated increases. In general, the plant is operated under pseudo steady-state conditions. It should also be mentioned that flow recovery by the RO system is 85%, therefore, the plant effluent flow rate is approximately 238,000 ± 10,600 m3/d. The average influent temperature is 32 ± 3.5 °C. Variations in temperatures were appreciable with two distinct seasons, summer and winter which reflects the climatic conditions prevailing in the state of Kuwait. It is also noted that the daily variations in flow rates coincide with those in temperature, with both being of a seasonal nature. The observed patterns of behavior are very coherent. Meanwhile, it has been observed that the effluent flow rate from the RO unit varied with temperature which is due principally to the change in viscosity of the water with change in temperature.

Figure 3

Variations in influent flow rate and temperature during the period 1 September 2010–31 August 2011.

Figure 3

Variations in influent flow rate and temperature during the period 1 September 2010–31 August 2011.

The pressure applied to the UF system is 3 bar whereas the pressure exerted on the RO system is in the range of 7–11 bar, 11–14 bar and 14–16 bar for the first stage, the second stage and the third stage of RO module's trains, respectively. The energy consumption across the RO unit in terms of water produced ranged from 1.4 to 1.7 kWh/m3. The RO membranes employed in this plant are low-pressure type membranes with the highest pressure applied being 16 bar (about 230 lb/inch2). It is noteworthy to mention that membranes used have been operated for almost 6 years with a replacement rate of only 10%. Biofouling of membranes was properly controlled as chloramine was dosed to the RO feed water to control the rate of biofouling.

System performance

The operation of the water reclamation plant (Figure 1) indicates that the influent to the UF system is the biologically treated wastewater after chlorination and micro-straining while the effluent of the UF system enters as influent to the RO system that produces the reclaimed water. Figure 4 displays three sets of data (UF system's influent, UF effluent/ RO influent and RO effluent) for each performance parameter. The performance data for each parameter were monitored over a period of 1 year. Monthly average data were calculated for each parameter and plotted in Figure 4 along with their standard deviation values calculated in each case.

Figure 4

Influent and effluent daily performance data: (a) pH, (b) TSS, (c) TDS and (d) BOD.

Figure 4

Influent and effluent daily performance data: (a) pH, (b) TSS, (c) TDS and (d) BOD.

The pH of influent to the UF system (Figure 4(a)) did not vary much over the 1-year period, being in the range of 6.9–7.1. It is important to stress that the UF effluent passes through a buffering well (Figure 1) so as to adjust the pH value to meet the RO system requirements. Therefore, the pH values as recorded directly prior to entering the RO system (6.6–6.8) were slightly lower than that of UF effluent. Meanwhile, the pH values of RO effluent increased again (7.1–7.3) as it passed through a stripper for removal of CO2 to adjust pH with a minimum amount of caustic before the distribution system. Ultimately, the effluent from RO system is chlorinated to be ready for water reuse.

The monthly average TSS values of UF influent were variable and ranged from 5 to 14 mg/L. The UF effluent TSS concentration levels decreased to a stable value of <1 mg/L and remained at this level in the RO system effluent (Figure 4(b)). This demonstrated that UF system is unique and efficient in removing TSS, thus preventing it from entering RO system and fouling the RO membranes (Peng et al. 2004). In contrast, the monthly average TDS values were more stable in the UF influent at 400–500 mg/L and remained almost unchanged in the UF effluent. Subsequently, the TDS values were greatly reduced to <10 mg/L and remained consistently at this level over the 1-year period (Figure 4(c)). For BOD5, the monthly average varied between 2 and 7 mg/L and decreased consistently to 1 mg/L in the UF system and further to <1 mg/L in the RO system effluent (Figure 4(d)).

The plots displayed in Figure 4 highlight the process efficiency by greatly reducing the concentration of each quality parameter and its standard deviation, as well as the process reliability by lowering the time variability of each parameter in the course of treatment. It should be noted that the relevance of dissolved organic matter and specifically trace or priority organics are the primary concern when using advanced processes for water reclamation. Such organics are better expressed as COD or TOC of water quality parameters. However, the data obtained from the studied plant do not include such parameters and are limited to BOD since the standards to be satisfied by the reclaimed water are expressed in terms of BOD as a measure for biodegradable organic matter. The TDS levels, conversely, account for such dissolved organic matter and their very low concentrations detected in the RO effluent reflect their retention by RO membranes.

The total coliforms (Figure 5) were highly variable in the UF system influent ranging between 0.2 × 106 and 10 × 106 colony-forming units (cfu)/100 mL and decreased to a range of 0.05 × 106 and 0.47 × 106 cfu/100 mL in the UF system effluent. It was further decreased substantially to a constant value of <1 cfu/100 mL (i.e. not detectable) in the RO system effluent. It is to be noted that reduction in total coliforms was expressed in percentages rather than in log units based on the data obtained from the plant. For instance, a bacterial removal of 3.0 log units (say from 105 to 102) designates a 99.9% removal.

Figure 5

Influent and effluent daily performance data for total coliforms.

Figure 5

Influent and effluent daily performance data for total coliforms.

For the UF system, the concentration-based removal efficiencies of TSS, TDS, BOD and total coliforms are 85 ± 7%, 10 ± 5%, 72 ± 13% and 85 ± 10%, respectively. Conversely, for the RO system, the corresponding removal efficiencies for each parameter are >95%. The overall removal efficiencies for the water reclamation plant combining UF and RO systems are >99% and the reclaimed water consistently complies with the quality standards as clearly demonstrated by the data in Figures 4 and 5, and as compared to various standards adopted in Kuwait as presented in Table 2. The quality of reclaimed water (Table 2) is comparable to that of desalinated water (drinking water) but at a much lower (one-third) treatment cost. The results obtained in this study agree very well with those reported by other researchers (Franks et al. 2004; Bartels et al. 2010; Stover 2013).

Table 2

Characteristics of wastewater and reclaimed water compared to Kuwait's standards

Parameter Unit Inflow to WWTP Inflow to water reclamation plant RO effluent for reuse Irrigation water standards Drinking water standards 
pH – 6.5–8 6.5–7.5 6–8 6.5–8 6.8–7.5 
Conductivity µs/cm 1,200–3,000 1,100–2,200 – 1,500 515 
TSS mg/L 100–500 <10 <1 15 – 
VSS mg/L 70–350 <7.0 <1 – – 
COD mg/L 250–750 <40 15 100 – 
BOD5 mg/L 100–400 <10 <1 20 – 
Grease and oil mg/L 10–50 NIL < 0.05 – 
TDS mg/L 700–1,800 800–1,500 < 100 – 400 
Chloride mg/L 200–400 200–400 – – 103 
Ammonia mg/L 15–50 1–5 <1 15 – 
Nitrite mg/L 0.04–0.7 0.1–1.5 <1 – – 
Heterotrophic plate counts cfu/100mL 2.40 × 109 103 – 
Total coliforms cfu/100 mL 3.20 × 108 400 400 
Faecal coliforms cfu/100 mL 4.10 × 107 0–10 20 
Salmonella cfu/100 mL 4.50 × 106 – 
Streptococci cfu/100 mL 1.40 × 107 – 
Fungi cfu/100 mL 2.10 × 105 2–100 – 
Parameter Unit Inflow to WWTP Inflow to water reclamation plant RO effluent for reuse Irrigation water standards Drinking water standards 
pH – 6.5–8 6.5–7.5 6–8 6.5–8 6.8–7.5 
Conductivity µs/cm 1,200–3,000 1,100–2,200 – 1,500 515 
TSS mg/L 100–500 <10 <1 15 – 
VSS mg/L 70–350 <7.0 <1 – – 
COD mg/L 250–750 <40 15 100 – 
BOD5 mg/L 100–400 <10 <1 20 – 
Grease and oil mg/L 10–50 NIL < 0.05 – 
TDS mg/L 700–1,800 800–1,500 < 100 – 400 
Chloride mg/L 200–400 200–400 – – 103 
Ammonia mg/L 15–50 1–5 <1 15 – 
Nitrite mg/L 0.04–0.7 0.1–1.5 <1 – – 
Heterotrophic plate counts cfu/100mL 2.40 × 109 103 – 
Total coliforms cfu/100 mL 3.20 × 108 400 400 
Faecal coliforms cfu/100 mL 4.10 × 107 0–10 20 
Salmonella cfu/100 mL 4.50 × 106 – 
Streptococci cfu/100 mL 1.40 × 107 – 
Fungi cfu/100 mL 2.10 × 105 2–100 – 

VSS, volatile suspended solids.

It should be mentioned that water quality parameters such as BOD and TSS are commonly measured by WWTPs since plant laboratories have cost-effective techniques for conducting such measurements, even at the very low concentrations found in the RO-treated effluents. Other parameters such as TOC, although more accurate to monitor removal of organics at low concentrations, require more costly equipment such as the carbon analyser which is not available in most treatment plant laboratories to be used on routine basis. Moreover, water quality standards often use parameters such as BOD, COD and TSS to determine suitability of treated wastewater effluents for reuse.

Frequency analysis

Cumulative frequency distributions of the influent and effluent concentrations for the parameters pH, TSS, TDS and BOD are shown in Figure 6. The 50% cumulative frequency is particularly important to express the variability of each parameter between 0.1 and 99.9% frequencies. For the pH, Figure 6(a) indicates that the pH of the UF influent was, for 50% of the time, ≤7.0 but decreased to a value of 6.8 in the UF effluent/RO influent and then increased to a value of 7.2 in the RO effluent. It is to be noted that the cumulative frequency plots are steep, indicating low variability in this parameter with time. In contrast, parameters such as TSS, TDS and BOD are removed as they pass through the membranes. For instance, TSS was ≤9 mg/L for 50% of time and decreased significantly to <1 mg/L in the UF effluent to protect the RO membranes, and remained at this level in the RO effluent (Figure 6(b)). This was also true for TDS which was ≤410 mg/L for 50% of time and remained almost unchanged in the UF effluent/RO influent where it was greatly reduced to only 1 mg/L in the RO effluent (Figure 6(c)). The BOD was ≤3.5 mg/L for 50% of time but was reduced down to 1.5 mg/L by UF and further to <1 mg/L in the RO effluent (Figure 6(d)).

Figure 6

Cumulative frequency distributions of the influent and effluent daily performance data: (a) pH, (b) TSS, (c) TDS and (d) BOD.

Figure 6

Cumulative frequency distributions of the influent and effluent daily performance data: (a) pH, (b) TSS, (c) TDS and (d) BOD.

Figure 7 shows that total coliform count was ≤3 × 105 cfu/100 mL for 50% of the time but was ≤1.3 × 105 cfu/100 mL for 99% of the time indicating high variability in this parameter. It is also decreased significantly to <1 cfu/100 mL in the UF effluent and remained so in the RO effluent with no variability. Figures 4 and 5 reveal that the UF system is effective in reducing the concentration level and variability of TSS and total coliforms, whereas the RO system is very effective in reducing the concentration level and variability of the TDS and BOD parameters. Together, the UF and RO systems complement each other in achieving almost complete removal of TSS, TDS, BOD and total coliforms in the water reclamation plant.

Figure 7

Cumulative frequency distributions of the influent and effluent daily performance data for total coliforms.

Figure 7

Cumulative frequency distributions of the influent and effluent daily performance data for total coliforms.

Frequency distribution plots obtained for all parameters studied (Figure 6) clearly highlight the performance reliability of the UF and RO systems. These plots, in addition to those displayed in Figure 4 for time variability of concentrations of the same parameters serve to verify the degree of conformity to the normal probability distribution and visualize the high process efficiency and performance reliability of the UF and RO systems used for water reclamation.

CONCLUSIONS

Based on the results obtained from statistical analyses of operating data it is possible to conclude that the performance of the water reclamation plant over a period of 1 year was not significantly influenced by variations in the wastewater flow rate and/or temperature. The data on process variables conform to a normal probability distribution. The study demonstrates that the overall removal efficiencies for the water reclamation plant combining UF and RO systems are >99% for TSS, TDS, BOD and total coliforms. Average removal efficiencies are the highest for TSS and total coliforms in the UF system, while they are the highest for TDS and BOD in the RO system. Cumulative frequency distribution analysis highlights process reliability indicating that UF and RO systems play an important role in consistently maintaining a stable plant performance and a high quality reclaimed water. The UF system proved essential for complementing successful performance of the RO system. The reclaimed water satisfies, at 99% frequency, the quality standards for potable water concerning TSS, TDS, BOD and total coliforms. The statistical analysis of performance data produces homogeneous groups of variables and signals the greater importance of cumulative frequency distribution in interpreting performance data.

ACKNOWLEDGEMENTS

The authors wish to thank the staff of the Sulaibiya WWTP from both Al-Kharafi Co. operating the biological treatment stage and General Electric (GE) Co. operating the water reclamation stage for providing necessary information and data.

REFERENCES

REFERENCES
Angelakis
A. N.
Bonoux
L.
Lazarova
V.
2003
Challenges and prospective for water recycling and reuse in EU countries
.
Water Sci. Technol: Water Supply
3
,
59
68
.
APHA
2005
Standard Methods for the Examination of Water and Wastewater
,
21st edn
.
American Public Health Association
,
Washington, DC
,
USA
.
Asano
T.
Burton
F. L.
Leverenz
H. L.
Tsuchihashi
R.
Tchobanoglous
G.
2007
Water Reuse: Issues, Technologies and Applications
.
McGraw-Hill
,
New York
,
USA
.
Avula
R. Y.
Nelson
H. M.
Singh
R. K.
2009
Recycling of poultry process wastewater by ultrafiltration
.
Innov. Food Sci. Emerg.
10
,
1
8
.
Bartels
C.
Franks
R.
Andes
K.
2010
Operational Performance and Optimization of RO Wastewater Treatment Plants
.
Technical paper
,
Hydranautics
,
Oceanside, CA
,
USA
.
Bódalo-Santoyo
A.
Gomez-Carrasco
J. L.
Gomez-Gomez
E.
Maximo-Martin
M. F.
Hidalgo-Montesinos
A. M.
2004
Spiral-wound membrane reverse osmosis and the treatment of industrial effluents
.
Desalination
160
,
151
158
.
Cheryan
M.
1986
Ultrafiltration Handbook
.
Technomic Publishing Company Inc
.,
USA
.
Dolar
D.
Perisa
M.
Kosutic
K.
Babic
S.
2013
NF/RO removal of enrofloxacin and its photodegradation products from water
.
Desalination Water Treat.
51
,
469
475
.
Fababuj-Roger
M.
Mendoza-Roca
J. A.
Galiana-Aleixandre
M. V.
Bes-Pia
A.
Cuaryas-Uribe
B.
Iborra-Clar
A.
2007
Reuse of tannery wastewaters by combination of ultrafiltration and reverse osmosis after a conventional physical-chemical treatment
.
Desalination
204
,
219
226
.
Franks
R.
Papukchiev
U.
Murkute
P.
2004
MF/RO membrane system for treatment of municipal secondary effluent wastewater
. In:
Proceedings of the Annual Water Reuse Conference
,
Phoenix, AZ
,
USA
.
Gagne
D.
2002
The world's largest membrane-based water reuse project
.
Technical Paper
,
Water & Process Technologies, General Electric
,
Terrace, PA
,
USA
.
Goosen
M. F. A.
Shayya
W. H.
1999
Water management, purification and conservation in arid climates
. In:
Water Management, Purification and Conservation in Arid Climates
(
Goosen
M. F. A.
Shayya
W. H.
, eds).
Technomic Lancaster
,
PA, USA
,
1, 1–6
.
Hamoda
M. F.
Brodersen
K. T.
Sourirajan
S.
1973
Organics removal by low-pressure reverse osmosis
.
J. Water Pollut. C. Fed.
45
,
2146
2154
.
Lazarova
V.
Asano
T.
Bahri
A.
Anderson
J.
2013
Milestone in Water Reuse: The Best
Success Stories.
International Water Association Publishing
,
London
,
UK
.
Lo
Y. M.
Cao
D.
Argin-Soysal
S.
Wang
J.
Hahm
T.
2005
Recovery of protein from poultry processing wastewater using membrane ultrafiltration
.
Bioresource Technol.
96
,
687
698
.
Murthy
G. V. P.
Choudhari
L. B.
2009
Treatment of distillery spent wash by combined UF and RO processes
.
Global NEST J.
11
,
235
240
.
Nelson
H. M.
2006
Performance of polymeric membrane systems in the treatment of poultry processing plant waste effluent. PhD Dissertation
,
University of Georgia
,
Athens, Georgia
,
USA
.
Stover
R.
2013
New high recovery reverse osmosis water treatment for industrial, agricultural and municipal applications
. In:
Paper IDAWC/TIAN13-429, presented at: The International Desalination Association World Congress on Desalination and Water Reuse, Tianjin, China
.
International Desalination Association (IDA), Topsfield, MA, USA
.
Tansel
B.
Regula
J.
Shalewitz
R.
1995
Treatment of fuel oil contaminated waters by ultrafiltration membranes
.
Desalination
102
,
301
311
.
WHO
2008
Guidelines for Drinking-water Quality
,
Report of World Health Organization
,
Vol. 1
,
3rd edn
.
World Health Organization
,
Geneva
,
Switzerland
.